一种改进的类电磁机制算法  

Improved electromagnetism-like mechanism algorithm

作　　者：姜建国[1] 刘永青[1] 尚海豹[1] 林龙军[1] 王双记[1,2] 

机构地区：[1]西安电子科技大学计算机学院 [2]中国人民解放军91388部队

出　　处：《西安电子科技大学学报》2013年第3期87-94,共8页Journal of Xidian University

基　　金：国家部委基础科研计划资助项目(D1120060967)

摘　　要：从类电磁机制算法的机理出发,研究了其寻优机制.针对原算法中存在的初始种群不均匀、局部搜索效率低、合力计算公式病态性严重等问题,提出了一种改进的类电磁机制算法.采用均匀设计方法产生初始种群;采用混合混沌-模式搜索进行局部搜索;运用改进后的公式计算粒子之间的作用力;设计了一种自适应移动公式来更新粒子.实验结果表明,改进后的算法具有更好的收敛效果和更高的求解精度.Based on the principle of the electromagnetism-like mechanism（EM） algorithm, the algorithm＇s optimization mechanism is studied. An improved electromagnetism-like mechanism algorithm is proposed to solve the problems of the original EM algorithm, such as non-uniform initial population, low search efficiency, and pathological feature in the total force formula. The new algorithm generates the initial population with the uniform design method, improves the local search algorithm with the hybrid chaotic pattern search method and calculates the total force between particles with the modified equation. Besides, an adaptive formula of movement is designed to update the locations of those particles. Experimental results show that the improved algorithm has a better convergence result and a higher solution accuracy.

关 键 词：类电磁机制算法 均匀设计 混沌搜索 模式搜索 全局优化 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]